Re: [SO-devel] GFF3 Specification

[Joachim B. <joa...@gm...>]

On 2013-02-26, at 1:32 PM, Peter Cock <p.j...@go...> wrote:
> ##sequence-region chr1 10000 12000
> ##sequence-region chr1 50000 52000
  I came across that syntax in Ensembl's GVF files recently. I find it rather confusing.

  As I understand it, "sequence-region" is saying something about the genomic landmark at hand. "chr1" here. Chromosome 1 has a well-defined start and stop coordinate for a particular genome build and those numbers should be used in the "sequence-region" pragma.

  Or putting it differently: What do you gain by chopping up "chr1" into arbitrary regions?

  The only application I see for non-continuous landmark coordinates would be to exclude regions that have not been sequenced. For example, around the centromere. But then, perhaps that could be expressed more explicitly in GFF3?

Best wishes,
Joachim